[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: DTD Notation raises a question
- From: Rod Davison <firstname.lastname@example.org>
- To: email@example.com
- Date: Wed, 11 Jul 2001 16:44:52 -0400
Ya know, there always seems to be a counterexample to the statement "the
different productions A and B are semantically equivalent".
In Linguistics, it is well known that true synonomy does not exist. Given
two putative semantic synonyms (like "medial" and "middle") -- one can always
find a discourse pragmatic usage of one where the other would not be allowed
("I am the middle child", *"I am the medial child.").
Since our production rules for XML define a limited form of a language,
perhaps it would useful to keep in mind that, like in languages, different
forms cannot be universally (i.e. in all contexts) synonymous. This may seem
like a rather whoo-whoo kind of statement but I guess it is motivated by the
thought that I have only seem determinism discussed in the context of
parsing, not in terms of generation.
The example provided by Deborah Aleyne Lapeyre below raises this question in
a very specific and easily resolvable setting. Specifically, two content
models which are initially seen as equivalent (applying the generalization
that extraneous parentheses do not change the model), but yet produce
different sets of "sentences". I
Anyone one done any work in this area or have any references? It is
something I am unfamiliar with and I'd love to read up on it.
On Wednesday 11 July 2001 03:45 pm, Deborah Aleyne Lapeyre wrote:
> >What's the difference between "ItemC*" and "(ItemC)*"?
> None, as everybody already said.
> Parens ALMOST never matter.
> That said, there is one warning. Since #PCDATA must
> be first in any grouping, there is a big difference
> (#PCDATA | itemC)* (valid)
> and the invalid
> ((#PCDATA | itemC))*
Rod Davison @ Critical Knowledge Systems Inc
"Historically speaking, the presence of wheels in Unix has never precluded
their reinvention." - Larry Wall